Valumigo

Artificial Analysis · 智能指数

定期获取并展示公开的基准测试资料（LMArena 用户投票排名、Epoch AI 测试分数、OpenRouter 使用量排名）。每个榜单均标注数据发布日期和获取日期。这些分数并非本站自行评定。

LMArena 分数来自用户并排查看两个模型的回答，再投票选出更好的一方（Elo 方式）。如果分数差异在置信区间内，可视为水平基本相近。

如何解读此排行榜

衡量什么： 这是 Artificial Analysis 直接运行多项评测后汇总得出的综合智能指数。数据通过 OpenRouter 公开数据 API 获取。

如何解读分数： 指数越高，表示在多项测试中整体表现越好。同一模型也会根据推理设置（Max、High 等）分别显示。

注意事项： Artificial Analysis 可能在不同版本中调整指数所包含的测试，因此请勿比较不同时期的得分。

本次排名的解读要点

  • 最高分为 Claude Opus 5.5 (Max, Default Fallback) 的 57.6。
  • 第 1 名与第 5 名的差距为 4.9。
  • 前 10 个模型中，Anthropic 的模型有 5 个，数量最多。

智能指数 Artificial Analysis

发布日期：2026-10-05 · 获取日期：2026-10-05

AnthropicClaude Opus 5.5 (Max, Default Fallback)
57.6
AnthropicClaude Sonnet 5.5 (Max, Default Fallback)
56.0
AnthropicClaude Fable 5.1 (Max, Default Fallback)
53.4
AlibabaQwen3.8 Max
53.4
OGPT-6 Astra (max)
52.7
OGPT-6.1 Sol (max)
51.8
AnthropicClaude Opus 5 (Max)
50.8
AnthropicClaude Fable 5 (Max, Opus 4.8 Fallback)
49.6
OGPT-6 Sol (Max)
47.6
OGPT-5.6 Sol (Max)
47.0
xGrok 4.7 (Xhigh)
46.4
XiaomiMiMo-V2.6-Pro
46.3
AlibabaQwen3.8 Max (0902)
45.4
ZGLM-5.3 (Max)
44.8
xGrok 4.6 (High)
44.3
Moonshot AIKimi K3 (Max)
43.6
OGPT-5.6 Terra (Max)
42.1
AnthropicClaude Opus 4.8 (Max)
41.8
ZGLM 5.3 Flash
41.8
iLing 3.1 Flash
41.1

图表仅显示同一模型各推理设置（high、max 等）中排名最高的一项。表格列出所有设置各自的排名。

查看表格（前 60 名）
排名模型开发方分数
1#1AnthropicClaude Opus 5.5 (Max, Default Fallback)Anthropic57.6
2#2AnthropicClaude Sonnet 5.5 (Max, Default Fallback)Anthropic56
3#3AnthropicClaude Fable 5.1 (Max, Default Fallback)Anthropic53.4
4#4AlibabaQwen3.8 MaxAlibaba53.4
5#5OGPT-6 Astra (max)OpenAI52.7
6#6OGPT-6.1 Sol (max)OpenAI51.8
7#7AnthropicClaude Opus 5 (Max)Anthropic50.8
8#8AnthropicClaude Fable 5 (Max, Opus 4.8 Fallback)Anthropic49.6
9#9OGPT-6 Sol (Max)OpenAI47.6
10#10OGPT-5.6 Sol (Max)OpenAI47
11#11xGrok 4.7 (Xhigh)xAI46.4
12#12XiaomiMiMo-V2.6-ProXiaomi46.3
13#13AlibabaQwen3.8 Max (0902)Alibaba45.4
14#14ZGLM-5.3 (Max)Z.ai44.8
15#15xGrok 4.6 (High)xAI44.3
16#16Moonshot AIKimi K3 (Max)Moonshot AI43.6
17#17OGPT-5.6 Terra (Max)OpenAI42.1
18#18AnthropicClaude Opus 4.8 (Max)Anthropic41.8
19#19ZGLM 5.3 FlashZ.ai41.8
20#20iLing 3.1 FlashinclusionAI41.1
21#21GoogleGemini 3.8 Flash (High)Google40.9
22#22AlibabaQwen3.8 2.4T A95BAlibaba39.9
23#23MetaMuse Spark 1.2 (Xhigh)Meta39.6
24#24DeepSeekDeepSeek V4.1 Flash (Reasoning, Max Effort)DeepSeek39.5
25#25GoogleGemini 3.7 Flash (High)Google39.1
26#26xGrok 4.5 (High)xAI38.8
27#27OGPT-5.5 (Xhigh)OpenAI38.4
28#28AnthropicClaude Sonnet 5 (Max)Anthropic38.2
29#29OGPT-6 Luna (Max)OpenAI38.1
30#30XiaomiMiMo-V2.6-FlashXiaomi37.9
31#31OGPT-5.6 Luna (Max)OpenAI37.3
32#32DeepSeekDeepSeek V4 Pro 0813 (Max)DeepSeek36
33#33DeepSeekDeepSeek V4 Flash 0731 (Max)DeepSeek34.3
34#34GoogleGemini 3.6 Flash (High)Google34
35#35MetaMuse Spark 1.1 (Xhigh)Meta33.7
36#36ZGLM-5.2 (Max)Z.ai33.7
37#37AlibabaQwen3.8 27B (Xhigh)Alibaba33.7
38#38GoogleGemini 3.5 Flash (High)Google32.6
39#39DeepSeekDeepSeek V4 Pro 0424 (Max)DeepSeek30.4
40#40AnthropicClaude Sonnet 4.6 (Max)Anthropic30.1
41#41GoogleGemini 3.1 Pro PreviewGoogle29.7
42#42AlibabaQwen3.7 MaxAlibaba29.5
43#43MiniMaxMiniMax-M3MiniMax29.2
44#44Moonshot AIKimi K2.6 (Reasoning)Moonshot AI27
45#45ZGLM-5.1 (Reasoning)Z.ai26.1
46#46XiaomiMiMo-V2.5-Pro (Reasoning)Xiaomi26
47#47Moonshot AIKimi K2.7 CodeMoonshot AI25.8
48#48TInkling SmallThinkingmachines25.7
49#49THy3Tencent25.3
50#50AlibabaQwen3.7 PlusAlibaba25.2
51#51TInkling (Xhigh)Thinkingmachines25
52#52xGrok 4.3 (High)xAI24.9
53#53iLing-3.0-flash-VLinclusionAI24.6
54#54DeepSeekDeepSeek V4 Flash 0420 (High)DeepSeek24.4
55#55OGPT-5.4 mini (Xhigh)OpenAI24.1
56#56USolar Mini 4Upstage24.1
57#57NVIDIANemotron 3 Ultra 550B A55B (Reasoning)NVIDIA22.9
58#58MiniMaxMiniMax-M2.7MiniMax22.8
59#59iLing-3.0-flash-FininclusionAI22.6
60#60GoogleGemini 3.5 Flash-LiteGoogle22.2

Source: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings). Licensed under CC BY 4.0.

数据：LMArena leaderboard dataset（CC BY 4.0）。修改：各榜单仅显示前 50 名，分数四舍五入。 https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset公司标志为各公司的商标，仅用于帮助区分（图标：Simple Icons）。数据：Epoch AI — Capabilities & benchmarking（CC BY 4.0）。 https://epoch.ai/benchmarksSource: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings). Source: OpenRouter evals (openrouter.ai) via OpenRouter (openrouter.ai/rankings). Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings). Licensed under CC BY 4.0. Source: OpenRouter (openrouter.ai/rankings), as of 2026-10-05. Licensed under CC BY 4.0. https://openrouter.ai/rankings